gsmiller commented on PR #12073:
URL: https://github.com/apache/lucene/pull/12073#issuecomment-1379404470

   @jpountz:
   > understanding why we're seeing a speedup with this change
   Good question. I was not familiar with `ReqExclBulkScorer`, but after taking 
a bit of a look, I have the same question now. I'll have to spend more time 
looking at these benchmark tasks to understand what might be going on here. 
Thank you for pointing this out.
   
   @rmuir:
   Thanks for the feedback! You raise some good questions here. My intuition 
was that reducing the size of the first phase match set would be beneficial in 
general. The reasoning being that, 1) it reduces the number of "candidates" 
that other potential conjunction clauses need to evaluate, and 2) it reduces 
the number of candidates that need to go to a second-phase (particularly if 
there are 2nd phase checks occurring _before_ the exclusion check). But I hear 
your point about term queries being a best-case scenario here. If the exclusion 
clauses are more complex, it's more difficult to reason about.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to